A Self-Repairing Execution Unit for Microprogrammed Processors

The emerging field of self-repair computing could have a major impact on deployable systems for space missions and defense applications. These systems must survive and perform at optimal functionality for long durations in unknown, harsh, and/or changing environments. Examples of such applications include outer solar system exploration, missions to comets and planets with severe environmental conditions, long-lasting space-borne surveillance platforms, defense systems, and monitoring and control of longterm nuclear waste and other hazardous environments. Self-repair computing could also greatly enrich commercial applications that require high availability and serviceability. These applications could range from biomedical devices to automotive applications. The proposed self-repair architecture for microprogrammed processors is transparent to the user and tolerates the occurrence of multiple faults in the device’s functional units. The processor achieves self-repair by letting the device dynamically reconfigure its internal microcode to execute required computations using only fault-free system units. One of the main novelties of our approach is that it does not require adding redundant or spare computational blocks to the system. The approach introduces a graceful degradation of device performance, but nevertheless lets the processor complete the requested operations even when multiple faults are present in its functional units. Researchers have done little work in the field of self-repair computing. Most studies focus on field-programmable gate arrays.